Bone-anchored hearing systems (BAHSs) are surgically implantable devices that bypass the outer and middle ear by directly stimulating the cochlea via a percutaneous implant anchored to the temporal bone (Tjellstrom and Hakansson 1995). BAHS have been commercially available for over 30 years and offered as a successful treatment for adults and children with conductive or mixed hearing losses, chronic middle ear disease, outer or middle ear malformations, and, more recently, in patients with single-sided deafness (Snik et al. 1995; Snik et al. 2004; Baguley et al. 2006). The efficacy of treatment with BAHS has been evaluated in several studies (for recent reviews, see Snik et al. 2019; Lagerkvist et al. 2020), especially to support the benefits of unilateral BAHS provision. Although the evidence supporting the benefits of bilateral fittings has been growing in the last decade (for reviews, see Janssen et al. 2012; Heath et al. 2022), the penetration of bilateral BAHS treatment is still rather low. A market survey showed that the penetration of bilateral fittings of BAHS in several European countries is about 6% (Cochlear bilateral market survey 2018), while bilateral hearing aid fitting is considered standard care and performed in more than 70% of all hearing aid users (Kochkin 2009; Abrams and Kihm 2015).
The fact that bilateral BAHS fitting is not as common in clinical practice as bilateral air conduction (AC) hearing aid fitting has been motivated for many years by cross hearing, that is, the extent to which stimulation via bone conduction (BC) excites the contralateral cochlea. Despite large interindividual differences, cross-hearing is very large at most frequencies. It has been shown that transcranial attenuation for stimulation at BAHS position is close to zero at most frequencies, except for frequencies between 2.5 and 5.7 kHz, where the median transcranial attenuation ranges from 2.5 to 7.0 dB (Stenfelt 2012). Thus, the sound transmitted via BC by one sound processor will excite both cochleae with only a small level difference at most frequencies. This has led to a common misconception that fitting BAHS unilaterally is sufficient since both cochleae will receive the input signal. However, binaural hearing relies on receiving the input signal at the two cochleae via two pathways that are acoustically separated. This makes it possible for the auditory system to compare the signal from the right cochlea with the one from the left cochlea and compute the difference in terms of time of arrival at the two ears and sound pressure level. These differences are referred to as interaural time and level differences (ITDs and ILDs) or binaural cues (Blauert 1997). When sound stimulation occurs via AC, the auditory system can mostly use ITDs for the low-frequency components of the incoming sound and ILDs for high-frequency components (Blauert 1997). For stimulation via BC, it has been suggested that ILDs may be the dominant binaural cue given that transcranial attenuation is greater at higher frequencies (Stenfelt and Zeitooni 2013). However, localization experiments using low-frequency versus high-frequency sounds suggest that both ITDs and ILDs are available binaural cues for BAHS users (Bosman et al. 2001; Priwin et al. 2004).
While the relative contribution of ITDs and ILDs for BAHS users remains somewhat unclear, it is increasingly recognized that patients with good (symmetrical) cochlear function can benefit from bilateral BC inputs (Janssen et al. 2012). Several studies have shown that bilaterally implanted BAHS users experience a benefit in binaural hearing in several perceptual measures: improved speech intelligibility in quiet and in noise, larger spatial release from masking, increased ability to localize sounds, and increased quality of life (van der Pouw et al. 1998; Bosman et al. 2001; Dutt et al. 2002; Stenfelt, 2005; Ho et al. 2009; Dun et al. 2010; Janssen et al. 2012; Stenfelt and Zeitooni 2013; Dun et al. 2013; Zeitooni et al. 2016). The benefit in binaural hearing obtained in these studies confirms that bilaterally implanted BAHS users can extract and process binaural cues, at least up to a certain extent (Stenfelt and Zeitooni 2013).
While a growing body of evidence highlights the advantages of bilateral BAHS treatment for binaural hearing, more research is needed towards understanding whether bilateral BAHS treatment may lead to additional benefits. The main goal of this study was to evaluate whether bilateral BAHS fitting can account only for benefits in binaural hearing or whether higher-level cognitive benefits are additionally involved, such as increased memory for speech. The reasoning behind this being that the incoming sound will be more precisely coded along the auditory system when both cochleae receive the signal via bilateral aids. Besides creating binaural cues, this may also affect cognitive functions, which have been shown to depend partly on accurate coding of timing information (Bartzokis 2004), by increasing the amount, speed, and synchronicity of information coming through the perceptual channel (Pichora-Fuller et al. 1995; Lunner 2003). The relationship between binaural and cognitive measures has already been explored in a few studies with normal hearing (Merat and Groeger 2003; Ries et al. 2010) and with hearing-impaired listeners (Neher et al. 2011; Neher et al. 2012) but, to the knowledge of the authors, no study has addressed this aspect for BAHS users.
A well-established measure of binaural benefit, the minimum audible angle (MAA) test, was first used to evaluate the benefits of bilateral implantation for binaural hearing in adult BAHS users (Mills 1958; Litovsky 1997; Litovsky et al. 2006). MAAs are a measure of spatial resolution (or localization acuity) that require the use of ITDs and ILDs for accurate left-right localization on the horizontal plane (Blauert 1997). One previous study compared performance in the MAA test for children that were either unilaterally fitted with one BAHS or bilaterally fitted with two BAHS (Dun et al. 2013). Dun et al. (2013) showed that the mean angle for correct right-left discrimination was 68° in the unilateral condition and 13° in the bilateral condition, hence showing a large bilateral benefit in localization acuity for children. However, the authors did not evaluate how different patient characteristics (e.g., type of hearing loss, pure-tone average, duration of hearing loss, age of BAHS implantation) affected performance in the MAA test, given the small sample of children that participated in the study. To the knowledge of the authors, the present study is the first to evaluate MAA thresholds in a group of adult BAHS users that were unilaterally fitted versus bilaterally fitted and to evaluate the relationship between MAA performance and patient characteristics (i.e., type of hearing loss, pure-tone average, age, and duration of bilateral implantation).
To evaluate higher-level cognitive benefits of bilateral implantation, a test of memory processing for speech, the Sentence-final Word Identification and Recall test (SWIR; Ng et al. 2013a; Ng et al. 2015) was used. Two previous studies used the SWIR test with BAHS and showed that when greater demands are imposed on the incoming signal, either because of the transmission system affecting sound quality (softband versus abutment; Lunner et al. 2016) or because of higher background noise (Westover et al. 2020), recall abilities deteriorated. These findings suggest that “high signal fidelity is important to avoid redirection of cognitive resources away from storage and to listening” (Lunner et al. 2016). Hence, the hypothesis of the present study was that bilateral BAHS users may need to allocate less resources to encode a speech signal that is transmitted via bilaterally fitted BAHS (as compared to encoding sound transmitted only unilaterally)—and, consequently, they can allocate more resources to remember it. The assumption behind this hypothesis was that listening to speech via bilaterally fitted BAHS is the natural everyday listening condition for the bilateral BAHS users included in this study—hence, they might achieve easier and faster encoding of a speech signal transmitted bilaterally than unilaterally. Because a trade-off occurs between the resources allocated for encoding the signal and those allocated for storing information (Rönnberg et al. 2008), we assumed that processing speech via bilaterally fitted BAHS would lead to more capacity available for storage of information and higher recall ability as compared to processing speech via unilaterally fitted BAHS.
Besides evaluating binaural and cognitive benefits of bilateral fitting in the laboratory, this study additionally investigated self-perceived performance in everyday life for bilaterally fitted BAHS users via the Speech Spatial and Quality of Hearing Scale (SSQ12; Noble et al. 2013). The strength of this study is to use three outcome measures (SWIR, MAA, and SSQ12) for the same bilaterally fitted listeners to evaluate how bilateral fitting affects the listeners’ performance in the laboratory and in daily life. Furthermore, this study evaluated the relationship between self-reported performance (obtained via the SSQ12 questionnaire) and bilateral performance as assessed in the laboratory (via the SWIR and MAA tests) to clarify whether psychoacoustical tests performed in the laboratory can reflect the listeners’ perceived benefit in daily life. In addition, the relationship between the three outcome measures and the listeners’ characteristics was also evaluated to better understand how duration of bilateral implantation, degree of hearing loss, and age modulate performance.
MATERIALS AND METHODS
A total of 29 listeners were screened for participation. There were three screening failures and two early terminations. Twenty-four listeners (12 females, 12 males) completed the study (mean age: 55 years old). All listeners were adult BAHS users (minimum 18 years and maximum 75 years), who were already bilaterally implanted with two BAHS and had been using both Ponto sound processors (Oticon Medical AB, Askim, Sweden) daily for at least 6 months. All listeners had a bilateral hearing loss, either conductive (N = 8) or mixed (N = 16). All listeners had BC thresholds within the fitting range of the test device (Ponto 3 SuperPower): pure-tone-average (PTA) BC, calculated at 0.5, 1, 2, and 3 kHz, on both sides lower or equal to 65 dB HL. All listeners had BC thresholds with an average difference in masked BC (at 0.5, 1, 2, and 4 kHz) between right and left ears within 10 dB. Twenty-one listeners had individual threshold differences within 15 dB, and, hence, were considered as having symmetric BC thresholds (Bosman et al. 2001), while three listeners had mildly asymmetric thresholds at one or two individual frequencies with more than 15 dB difference (two listeners, 20 dB; one listener, 30 dB). Figure 1 shows the mean and individual hearing thresholds for right and left ears (AC; BC). The listeners' demographics and other baseline characteristics are reported in Table 1.
TABLE 1. -
Patient demographics and baseline characteristics
||Mean = 54.96 (SD = ± 13.67)
Median = 60.5(min: 23; max: 74)
||n = 12 (50%)
||n = 12 (50%)
|Type of hearing loss
||n = 8 (33.33%)
||n = 16 (66.67%)
|PTA AC right (dB HL)
||Mean PTAAC = 64.79 (SD = ±13.81)
Median PTAAC = 64.38 (36.25; 97.5)
|PTA AC left (dB HL)
||Mean PTAAC = 66.88 (SD = ±13.81)
Median PTAAC = 68.75 (41.25; 105)
|PTA BC right masked (dB HL)
||Mean PTABC = 33.02 (SD = ±11.48)
Median PTABC = 32.5 (7.5; 55)
|PTA BC left masked (dB HL)
||Mean PTABC = 33.49 (SD = ±10.85)
Median PTABC = 33.75 (11.25; 53.75)
||n = 17 (70.83%)
| Non-sequential (simultaneous)
|Time in-between surgeries for sequential implantations (yrs)
||Mean = 7.16 (SD = ± 5.29)
Median = 6.54(min: 0.58; max: 18.83)
|Current (own) sound processors
||Ponto 3: n = 5 (20.83%)
Ponto 3 Power: n = 5 (20.83%)
Ponto 3 SuperPower: n = 1 (4.17%)
Ponto Plus: n = 8 (33.33%)
Ponto Plus Power: n = 5 (20.83%)
||Ponto 3: n = 6 (25%)
Ponto 3 Power: n = 4 (16.67%)
Ponto 3 SuperPower: n = 1 (4.17%)
Ponto Plus: n = 8 (33.33%)
Ponto Plus Power: n = 5 (20.83%)
PTA AC and PTA BC are calculated as the average threshold at 0.5, 1, 2, and 4 kHz. The distribution of continuous variables is given as mean, SD, median, minimum and maximum (mean [SD]/median [min; max]). The distribution of categorical and dichotomous variables is given as number and percentage (n (%)).
AC, air conduction; BC, bone conduction; PTA, pure-tone-average.
The study was approved by the Research Ethics Committee and NHS Health Research Authority (HRA) London - City & East (REC Reference 19/LO/0355).
The study included two visits per listener, each of up to 2 hours in duration.
At the first visit, audiometry was performed (AC, BC), followed by bilateral and unilateral fitting of the test device(s) (Ponto 3 SuperPower). The gain prescription was done using the standardized NAL-NL1 prescription for BAHS as implemented in Genie Medical BAHS (Genie Medical BAHS, Version 16.1). The gain prescription takes the BC in-situ threshold (unmasked) for each ear into account, and thereby reduces any intercranial asymmetries. This gain prescription is, like for all hearing aid rationales, based on a generalized loudness model. However, no control was performed for interaural symmetry on an individual patient basis. Hence, the data collected in this study reflect the situation for BAHS patients fitted with the methods used in clinical practice. The prescription of gain provided 3 dB more gain for the unilateral fitting as compared to the bilateral fitting (as a compensation for binaural summation). Directionality was set to omnidirectional mode and noise reduction was turned off. No fine-tuning was performed. After fitting was completed, the MAA test (Litovsky 1997; Dun et al. 2013) was performed.
Between Visit 1 and Visit 2
The listeners were asked to complete the SSQ12 questionnaire at home before the second study visit, regarding their everyday performance with their own BAHS sound processors. All participants were bilaterally fitted and used two BAHS in their everyday life.
At the second visit, the listeners performed the SWIR test (following the procedure described in Lunner et al. 2016). The details and set up for each test are explained in the following sections.
Blinding the Listeners to the Test Conditions
In a best attempt to blind the listeners to the test conditions, two bone-anchored sound processors were worn at the same time for both the unilateral and the bilateral conditions (for both the MAA and SWIR test). In the bilateral condition, both Ponto 3 SuperPower were fitted. In the unilateral condition, whilst the listener wore two devices, only one active Ponto 3 SuperPower was fitted on the selected ear (the second inactive device was muted before being mounted on the listener’s abutment). Hence, four different sound processors were used in total for each listener (two for the bilateral condition and two for the unilateral condition). The experimenter had to change the sound processors when there was a change in condition (bilateral versus unilateral). The test ear chosen for the unilateral condition was balanced across listeners: listeners with an uneven identification number were tested on their left ear, and listeners with an even identification number were tested on their right ear. During testing of the unilateral condition, the unaided ear was not occluded (e.g., via ear plugs) nor masked to ensure testing was reflective of a realistic scenario.
The study was conducted in a sound-insulated audiology room at the Queen Elizabeth Hospital (Birmingham, United Kingdom). The University Hospitals Birmingham is one of UK’s pioneering centers for BAHS and has since the late 1980s implanted more than 2500 recipients with over 500 of these receiving bilateral provision.
Spatial Resolution: MAA Test
A similar test method as the one used by Dun et al. (2013) was used here. The test started with two loudspeakers positioned at the left- and right-hand side of the listener at a distance of 1 m and at an azimuth angle of ±90°. One sound stimulus consisting of broadband pink noise (500 ms duration) was presented for each trial either from the right or the left loudspeaker. Pink noise has less energy in the high frequencies (where spectral cues are mostly extracted) and it was, thus, chosen to minimize the use of monaural spectral cues. The listeners were instructed to not turn their head and look at a sign positioned at 0° azimuth for the duration of the experiment. The listeners had to report the direction from which they thought the sound came from (by saying “Right” or “Left” after each stimulus presentation) and to guess if they did not know. After 10 stimulus presentations (trials), the loudspeakers were repositioned at decreasingly smaller angles. The measured angles were ±90°, ±60°, ±30°, ±15°, ±10°, and ±5° (i.e., 90°, 60°, 30°, 15°, 10°, and 5° to the left- and right-hand side of the listener relative to 0° azimuth). Signs on the floor indicated the exact angle positions. For each angle configuration, the number of correctly identified noise bursts (out of 10) was reported in terms of percent correct. Sound levels were randomized between 60-, 65-, and 70-dB SPL across the 10 trials (each level was presented a similar amount of times) to prevent the listeners from comparing the stimulus sound level across trials and using it as a monaural cue for sound source localization. The order of the conditions was balanced; listeners with an even identification number performed the unilateral condition first (10 trials for each of the six angle configurations), and listeners with an odd identification number performed the bilateral condition first (10 trials for each of the six angle configurations).
For each listener and condition, the percentage of correct responses was plotted as a function of the measured angles and a psychometric function was fitted to the percent correct. The MAA for each listener was calculated as the angle (in degrees azimuth, °) corresponding to 80% correct performance on the fitted psychometric function. If the fitted psychometric function was lower than 80% correct at all angles, the MAA was reported as “Not Measurable (NM)” (Litovsky et al. 2006). To include this datapoint in the statistics, a NM MAA was conservatively approximated to 90°, i.e., to the largest possible angle. Because the test was designed as a two-alternative-forced choice (“Right” or “Left”) procedure, the fitted psychometric function was set to start at the 50% chance level for a 0° azimuth angle (Levitt 1971). Because the MAA was obtained from a psychometric function that started at an abscissa of 0°, it was possible to obtain a MAA smaller than 5°.
Memory for Speech: SWIR Test Procedure
The SWIR test (Ng et al. 2013a; Ng et al. 2015) is a test for memory processing of speech, that aims to evaluate the participant’s recall ability when speech is presented at high intelligibility levels. The test was conducted at a fixed speech level (70 dB SPL), while the noise level was adjusted for each participant to achieve 95% correct in last-word identification. In other words, the intelligibility was adjusted to be almost at ceiling such that the memory task could be performed when about six to seven last words were intelligible for each list of seven sentences. Individually adjusting the noise level for each participant requires to carry out an adaptive procedure before conducting the SWIR test, namely a speech-in-noise test followed by a SWIR training (Ng et al. 2015; Lunner et al. 2016). The methods for the speech-in-noise test, the SWIR training, and the SWIR test are explained in the following paragraphs. All the above-mentioned tests were conducted with target sentences presented from the front (at 0° azimuth) and a four-talker babble noise coming from the side and back of the listener (loudspeakers positioned at ±90° and ±135° azimuth).
Three lists of 16 sentences were presented to the participant. The speech material used was the Bamford-Kowal-Bench Sentence Test (Bench et al. 1979) with a male target speaker. The listener’s task was to listen and repeat each sentence. The test was carried out at a fixed speech level of 70 dB SPL (C), while the masker level was adaptively varied to obtain the signal-to-noise ratio (SNR) corresponding to 80% correct recognition (speech recognition thresholds, SRT80) for each listener. The best (lowest) SRT80 obtained in the last two lists was chosen as the starting level for the SWIR training. The speech-in-noise test was performed only for the unilateral condition (i.e., the most demanding condition).
The SWIR training has two purposes, on one hand, to train the participants for the SWIR task and, on the other hand, to adjust the masker level for each participant. Specifically, the masker level during the SWIR training is adjusted from 80% correct sentence recognition (obtained during the speech-in-noise test) to 95% last-word identification. The procedure to adjust the masker level is described below (Lunner et al. 2016). Four lists of seven sentences (Institute of Hearing Research, IHR; Macleod and Summerfield 1990) were presented to the participant (see an example of a list in Supplemental Appendix 1, Supplemental Digital Content 1, https://links.lww.com/EANDH/B70). The IHR sentences have a similar syntax and vocabulary as the Bamford-Kowal-Bench sentences and are recorded by the same male speaker. The listener had two tasks: repeat the last word immediately after listening to each sentence (repetition task) and recall as many last words as possible, in any order, at the end of each list of sentences (recall task). The SWIR training was conducted at a fixed speech level of 70 dB SPL (C), while the masker level was adaptably varied (Ng et al. 2015; Lunner et al. 2016). The starting level of the masker was set to the level obtained for the best (lowest) SRT80 from the speech-in-noise test. The masker level was then adjusted after each list depending on how many last words were correctly repeated according to the following scoring rules: the masker level was decreased by 2 dB if zero to three last words were correctly repeated, by 1 dB if four or five words were correctly repeated, and it was left unchanged if six or seven words were correctly repeated. Hence, after four lists, the masker reached a level to achieve 95% correct in the repetition of the last words (Ng et al. 2015; Lunner et al. 2016). The SWIR training was performed only for the unilateral condition (i.e., for the most demanding condition).
Five lists of seven IHR sentences (see an example of a list in Supplemental Appendix 1, Supplemental Digital Content 1, https://links.lww.com/EANDH/B70) were presented to the participant for each of the two conditions, unilateral and bilateral, that is, 10 lists were presented in total. The listener had two tasks: repeat the last word immediately after listening to each sentence (repetition task) and recall as many last words as possible, in any order, at the end of each list of sentences (recall task). The SWIR test was conducted at a fixed speech level of 70 dB SPL (C) and at a fixed masker level, as obtained for each participant during the SWIR training. Hence, speech and masker levels were fixed for each participant to achieve almost full intelligibility, that is, 95% correct in last-word identification. The aim was to evaluate the listeners' recall ability in the unilateral and bilateral conditions. The order of which the two conditions (unilateral and bilateral) were tested in the SWIR test was randomized. The order of the lists was also randomized across listeners.
For each listener, the percentage of total recall in the SWIR test was calculated as the number of correctly recalled words out of the number of repeated words per list (for details, see Ng et al. 2013a). Note that if a word was initially repeated incorrectly, it could still be recalled correctly and hence it can be scored as correctly recalled (Ng et al. 2013a). For each listener, the total recall for unilateral and bilateral conditions was calculated as the average of total recall across all five lists tested for each condition. The percentage of correctly recalled words was also calculated for each serial position, that is, the position of the sentence in the list (see Supplemental Appendix 1, Supplemental Digital Content 1, https://links.lww.com/EANDH/B70). The first to second, third to fifth, and sixth to seventh sentences were partitioned into the primacy, asymptote, and recency positions, respectively (Ng et al. 2015; Lunner et al. 2016). Hence, the percentage of correctly recalled words (out of the number of repeated words) was also calculated for the primacy, asymptote, and recency positions for each listener.
Self-Reported Performance: SSQ12
A reduced version of the SSQ questionnaire (Speech, Spatial and Qualities of Hearing scale), the SSQ12, was used in this study. The SSQ12 was developed by Noble et al. (2013) and contains 12 items that reflect the perceived performance in speech intelligibility, spatial abilities, and sound quality.
Variables were tested for normality using the Shapiro–Wilk test. Paired t tests were performed where the data was normally distributed. Wilcoxon signed-rank tests were performed where the data was not normally distributed. In case of more than one factor involved in the design, mixed-linear models were performed with listener as a random factor. Post-hoc analysis was performed via contrasts of least-square means: one-tailed planned comparison for contrast related to the hypothesis of better performance in the bilateral versus unilateral condition or two-tailed comparisons for contrast not related to the hypothesis. Tukey method was used to correct for multiple comparisons. The decision of the hypothesis test was done at the 5% significance level. Pearson’s correlations (or Spearman’s where the data were not normally distributed) were used to test correlations. In case of multiple correlations, Bonferroni correction was applied (where appropriate) to correct for multiple comparisons. The data were analyzed using MATLAB (MathWorks, Natick, MA). Mixed-linear models were implemented in R-studio using the statistical package lmerTest (Kuznetsova et al. 2017).
Spatial Resolution: MAA Test
The left panel in Figure 2 shows the mean percentage of correct responses in the right/left discrimination task (MAA test), with the fitted psychometric functions for each condition. The results show that the mean performance in the bilateral condition was above 80% correct at all angles, while the mean performance in the unilateral condition was below 80% correct at all angles. In the unilateral condition, there was, as expected, a bias towards the aided side (Van Wanrooij and Van Opstal 2007); on average, 75% of the responses were reported at the aided side independently of the actual stimulus location.
The percentage of correct responses was analyzed with a mixed-model ANOVA with condition (unilateral, bilateral) and angles as fixed factors and listener as random factor. The analysis showed significant main effects of condition [F(1, 253) = 198.95; p < 0.0001] and angles [F(5, 253) = 7.44; p < 0.0001] and no significant interaction of condition and angles [F(5, 253) = 0.48; p = 0.794]. Hence, the difference in performance between unilateral and bilateral conditions, which was of at least 20% points at each angle, was significant across the tested angles.
The right panel in Figure 2 shows the MAA (i.e., the angle to obtain 80% correct) obtained in the unilateral and bilateral conditions. Performance in the right/left discrimination task improved significantly from a median MAA of 75.04° in the unilateral condition to 3.61° in the bilateral condition (paired one-tailed Wilcoxon signed-rank test, p < 0.0001). The average improvement in performance was 54.28°.
Note that 50% of the listeners (12 out of 24) were unable to achieve 80% correct at any angle in the unilateral condition and, hence, their MAA in the unilateral condition was NM. The MAA obtained in the unilateral condition was affected by the AC thresholds of the unaided side (which determined the availability of binaural cues). In fact, all listeners with a NM MAA in the unilateral condition had PTA AC thresholds (across frequency) at the unaided side greater than 60 dB HL. No listener with PTA AC thresholds greater than 80 dB HL obtained a measurable MAA in the unilateral condition. In contrast, only one listener had a NM MAA in the bilateral condition. Thus, the improvement of bilateral fitting for 11 listeners out of 24 (46%) was from “not being able to discriminate right versus left” to “being able to discriminate right versus left” (at 80% correct). In addition, there was no correlation between individual MAAs in the bilateral condition and PTA BC (mean across all frequencies averaged for Right and Left ear; Spearman correlation: ρ = 0.27; p = 0.196). Hence, performance in the MAA test (bilateral condition) did not depend on whether the hearing loss was conductive or mixed.
Memory for Speech: SWIR Test
The SWIR test was carried out at an individually adjusted SNR (obtained from the SWIR training), which was, on average, 4.9 dB (SD: ±4.2 dB). This SNR led to the following mean performance in the repetition task: 96.99% (SD: ± 4.17%) in the unilateral condition and 96.90% (SD: ± 3.25%) in the bilateral condition. In other words, intelligibility was adjusted to be close to 100% to be able to compare performance in the recall task across the two conditions. The mean total recall (across listeners) in the SWIR test was 55.03% in the unilateral condition and 57.23% in the bilateral condition.
The mean percentage (across listeners) of correctly recalled words for each condition and serial position (primacy, asymptote, recency) is presented in Figure 3. The difference between unilateral and bilateral conditions was rather small and around 1 to 2% points for sentences in primacy (recall unilateral: 47.7%; recall bilateral: 48.8%) and recency (recall unilateral: 85.5%; recall bilateral: 83.3%) positions, while it was above 6% points for sentences in asymptote position (recall unilateral: 39%; recall bilateral: 45.4%).
The results for each serial position were analyzed with a mixed-model ANOVA with condition (unilateral, bilateral), test repetition (i.e., the presentation of 10 lists), and serial position (primacy, asymptote, recency) as fixed factors and listener as random factor. The analysis showed a significant main effect of serial position [F(2, 663.99) = 136.37; p < 0.0001] and a significant interaction between condition and repetition [F(4, 664.08 = 2.44; p = 0.046)]. A two-tailed comparison to evaluate the effect of serial position showed that the recall in the recency position was significantly higher than in the primacy and asymptote positions (p < 0.0001). A one-tailed planned comparison was performed to evaluate the effect of condition for each repetition. Significantly higher performance in the bilateral condition was obtained for the second repetition/list (10.2% difference; p = 0.022).
Overall, the fitting of bilateral BAHS failed to demonstrate a significant effect on memory for speech (the main effect of condition was not significant). However, a significant interaction between condition and repetition was observed, revealing a significantly higher performance in the bilateral condition than in the unilateral condition for one out of the five repetitions (lists).
In addition, there was no correlation between recall performance in the unilateral condition and PTA AC on the unaided side and between recall performance in the bilateral condition and PTA BC, indicating that audibility was not a factor affecting performance (because intelligibility was adjusted to be close to 100% correct).
Self-Reported Performance: SSQ12
The results of the SSQ12 questionnaire are shown in Figure 4. The 12 questions of the questionnaire (Q1 to Q12) are listed in Supplemental Appendix 2, Supplemental Digital Content 1, https://links.lww.com/EANDH/B70. A higher score indicates better self-reported performance (a score of 0 corresponds to “Not at all” and a score of 10 corresponds to “Perfectly”). The listeners could also mark the question as “Not applicable”. Note that the obtained scores in the SSQ12 refer to the perceived performance in everyday life with own sound processors—which for most users were Ponto Plus (Power) or Ponto 3 (Power).
Self-reported performance obtained in the SSQ12 questionnaire was, on average, 4.5 (SD = 2.1). Self-reported performance for each scale was, on average, 4.4 for speech (SD = 2.3), 3.7 for spatial (SD = 2.5), and 5.1 for qualities of hearing (SD = 2.3).
Performance in SWIR, MAA, and SSQ Versus Age
Figure 5 (top panels) shows the relationship between recall performance and age, for each sentence serial position (primary, asymptote, and recency). When combining unilateral and bilateral conditions, a significant negative correlation was obtained between recall performance for the asymptote serial position and age (Spearman correlation: ρ = −0.37; p = 0.011), but not for primacy (Spearman correlation: ρ = −0.27; p = 0.061) and recency (Spearman correlation: ρ = 0.07; p = 0.627). The correlation for the asymptote position was significant also after Bonferroni correction for multiple comparisons (n = 3), p < 0.0167.
Because performance in the recall task decreased significantly with age, the benefit in recall performance (i.e., the difference in recall between bilateral and unilateral conditions) was also calculated for all listeners up to 65 years old (N = 20) and listeners above 65 years old (N = 4). Although these two age groups are not balanced for number of listeners, 65 years old was chosen since this was the maximum age of the listeners included in previous studies using SWIR (Ng et al. 2013; Ng et al. 2015). Figure 5 (bottom panels) shows the benefit in recall performance for each sentence serial position (primary, asymptote, and recency) for the two age groups. A significant benefit of 9.7% points (i.e., higher recall in the bilateral than unilateral condition) was obtained for the asymptote serial position in the group of listeners up to 65 years old (Wilcoxon sign-rank test: p = 0.007). This benefit was significant also after Bonferroni correction for multiple comparisons (n = 6), p < 0.0083. No significant benefit in recall performance was obtained for the other serial positions (p > 0.05).
There were no significant correlations between bilateral benefit obtained in the MAA test and age (Spearman correlation: ρ = −0.09; p = 0.691) and between total SSQ12 scores and age (Spearman correlation: ρ = −0.21; p = 0.318).
Performance in SWIR, MAA, and SSQ for Sequential Versus Simultaneous Implantation
This study was not specifically designed to evaluate differences in bilateral benefit between listeners who underwent sequential implantation (i.e., right and left implantation performed at different dates) versus simultaneous implantation (i.e., right and left implantation performed during the same surgery). However, this was also evaluated when investigating the relationship between listeners’ characteristics and the three outcome measures.
The benefit in recall performance (i.e., the difference in total recall between bilateral and unilateral condition) observed in the simultaneously implanted listeners (N = 7) was significant (mean benefit = 5.3% points; SD = ±3.4; paired two-sided Wilcoxon sign-rank test; p = 0.016), while no significant difference was observed in the sequentially implanted listeners (N = 17; mean benefit = 0.9% points; SD = ±7.6; paired two-sided Wilcoxon sign-rank test p = 0.585). There was, however, an age bias: all listeners above 65 years were sequentially implanted. Hence, the lack of a significant benefit in recall performance for the sequentially implanted listeners might have been limited by the older age of the listeners in this group (mean age sequentially implanted = 57.1 years ± 13.6 years; mean age simultaneously implanted = 49.7 years ± 13.4 years).
There was no significant difference in the bilateral benefit obtained in the MAA test for simultaneously versus sequentially implanted listeners (p = 0.694), nor in the total SSQ12 scores (p = 0.394).
Bilateral Benefit and Duration of Bilateral Implantation
The bilateral benefit obtained in the MAA test (difference in MAA between unilateral and bilateral conditions) was not correlated with duration of bilateral implantation (years after implantation with two BAHS; Spearman correlation: ρ = 0.21; p = 0.314). The bilateral benefit obtained in the SWIR test (difference in percent total recall between bilateral and unilateral conditions) was also not correlated with duration of bilateral implantation (Spearman correlation: ρ = −0.11; p = 0.6).
Self-Reported Bilateral Performance Versus MAA and SWIR
The relationship between self-reported performance in everyday life with two BAHS and the obtained performance in the laboratory with two BAHS was evaluated, for both the MAA and SWIR test.
Figure 6 shows the SSQ12 scores in the spatial subscale (questions six to eight) versus the MAA in the bilateral condition. There was no correlation between the two measures (Spearman correlation: ρ = 0.15; p = 0.47). Self-reported performance varied greatly in the spatial-related questions of the SSQ12, even for listeners that achieved similar performance in the MAA test.
Figure 7 shows the individual SSQ12 scores (left panel: speech subscale; mid panel: spatial subscale; right panel: qualities subscale) versus performance in total recall in the SWIR test for the bilateral condition. There was a significant correlation between total SSQ scores and recall performance (ρ = 0.56; p = 0.004), as well as for speech and qualities subscales (Spearman’s correlation coefficients and the p-values are reported in each panel of Fig. 7). The correlations were significant also after Bonferroni correction for multiple comparisons (n = 4), p < 0.0125.
Bilateral Fitting and Benefits for Localization Acuity
Performance in localization acuity was poor in the unilateral condition: on average, 75% of the responses were reported towards the aided side and 50% of the listeners could not discriminate whether a sound came from the right or left at any angle (NM MAA). These listeners with a NM MAA had AC thresholds at the unaided side greater than 60 dB HL (PTA), meaning that the signal was mostly transmitted via BC from the aided side and, hence, binaural cues were, to a large extent, not available. The fact that there was a large bias towards the aided side in the unilateral condition indicates that when listeners with a bilateral conductive or mixed hearing loss have symmetric BC thresholds and are only unilaterally aided with a BAHS, the cochlea ipsilateral to the BAHS will receive the signal at a higher level than the contralateral cochlea (at least at higher frequencies), leading the listener to bias the location to the aided ear (Van Wanrooij and Van Opstal 2007). It should be noted that performance in the unilateral condition represents an acute condition, because the participants of the present study are not used to listening with only one device. Hence, performance in the unilateral condition of the MAA test might differ for listeners who are unilaterally fitted with one BAHS in everyday life. Five listeners performed relatively well in the unilateral condition (MAAs lower than 8°), presumably because their AC thresholds at the unaided side were not too large, especially around 3 kHz (i.e., AC thresholds between 35 and 55 dB HL). Receiving bilateral sound transmission in a frequency range where transcranial attenuation is at its largest allowed these few listeners to use some binaural cues even when unilaterally fitted.
Performance in localization acuity improved significantly in the bilateral condition. The smallest angle to discriminate right versus left (at 80% correct) decreased from 75.04° in the unilateral condition to 3.61° in the bilateral condition. The large improvement in performance indicates that bilaterally fitted BAHS users can extract and use binaural cues (ILDs and/or ITDs) to localize sounds on the horizontal plane, as similarly shown for children by Dun et al. (2013). Although each BAHS transmits the sound to both cochleae, a degree of transcranial attenuation occurs at higher frequencies (Stenfelt 2012), which enables some acoustic separation of the two cochleae—at least for listeners with symmetrical BC thresholds—and makes it possible to extract and process binaural cues to some extent (Stenfelt and Zeitooni 2013). Performance in the bilateral condition was not dependent on PTA BC (and it was independent of whether the hearing loss was conductive or mixed) and the benefit of bilateral versus unilateral condition was also not related to the duration of bilateral implantation.
Although the use of monaural cues in the MAA test cannot be completely ruled out, it should be noted that the test was designed to minimize the use of monaural cues for localization. Monaural level cues were minimized by using three different sound levels for stimulus presentation (Bernstein 2004; Dun et al. 2013). Hence, the listeners could not discriminate right versus left based on a monaural level cue (by comparing the level of the current stimulus with the previous one). In addition, monaural spectral cues were reduced by using pink noise (instead of white noise) which has less energy in the high frequencies (where spectral cues are mostly extracted). Monaural spectral cues, which are caused by reflections from pinna, head, and torso, are generally helpful to localize sounds on the vertical and not on the horizontal plane (Hofman and Van Opstal 1998). Patients with single-sided deafness (profound unilateral sensorineural hearing loss) may learn to use monaural spectral cues also for localization on the horizontal plane since they do not have access to binaural cues (Agterberg et al. 2019). However, it is unlikely that the bilateral BAHS users that were included in this study made use of monaural spectral cues in the MAA test. In fact, bilaterally fitted BAHS users can use ITDs and ILDs (Bosman et al. 2001; Priwin et al. 2004) so, unlike SSD patients, they are not trained to primarily use monaural spectral cues for localization on the horizontal plane. This is further supported by the poor localization performance obtained in the unilateral condition. If the participants of the present study would have been able to use monaural spectral cues, they would have performed better in the unilateral condition.
Hence, the results obtained here demonstrate that BAHS users with a symmetrical hearing loss (conductive or mixed) can highly benefit from bilateral fitting and achieve a spatial resolution performance that is rather accurate. Previous studies have shown that MAA thresholds in normal-hearing children are about 12 to 19° at 6 mo (Ashmead et al. 1987), decrease to 4 to 6° by 18 mo, and to 1 to 2° by 5 years, when they are not significantly different from adult MAAs (Litovsky 1997). The findings of this study show that bilaterally fitted BAHS users with a symmetrical hearing loss can make good use of binaural cues to reach a median MAA of 3.61°—hence, a performance that is not largely different from normal-hearing performance. It should be mentioned that gain was prescribed according to the unmasked BC thresholds for each ear based on a standardized rationale, and therefore any intercranial asymmetries were reduced, but not necessarily completely absent. The data obtained here reflects right-left discrimination performance when using the prescription methods as in clinical practice.
Finally, it should be noted that the MAA test is a test of spatial resolution (or localization acuity), where the listeners need to discriminate between right or left source location. A previous study has shown a moderate correlation between right-left discrimination performance (i.e., MAA) and sound source localization in a multi-source task (Grieco-Calub and Litovsky 2010). The authors showed that the MAAs were much smaller than the angle measured in the localization task: for MAAs smaller than 20°, performance in the localization task was highly variable with deviation errors from target location ranging between 19.1° and 44.1°. In addition, the authors showed that poor performance in the right-left discrimination task was associated with poor performance in the localization task. Hence, an accurate performance in a right-left discrimination task does not necessarily predict high localization accuracy in a multi-source task, but it has rather been suggested that good performance in the MAA test may be a prerequisite for sound source localization (Grieco-Calub and Litovsky 2010). A big advantage of the MAA test over a test of sound source localization is that it only requires two loudspeakers and hence it is much easier to perform in a clinical setting.
Bilateral Fitting and Memory for Speech
In the SWIR test, each listener was tested at an individually adjusted SNR. On average, the SNR was 4.9 dB and listeners could repeat around 97% of the final words in both the unilateral and bilateral conditions. This ensured that recall performance was not dependent on audibility because intelligibility performance was fixed at a very high level for all listeners. In addition, a positive SNR of 4.9 dB is representative of an everyday listening environment for hearing-impaired listeners and, hence, considered ecologically valid (Smeds et al. 2015).
Performance in the recall task did not differ significantly between unilateral and bilateral conditions (i.e., the main effect of condition was not significant), and the average difference in recall performance between the two conditions was 2.19% points. This finding suggests that bilateral fitting did not have a significant effect on auditory memory for speech. However, there was a significant interaction between condition and repetition, revealing a significantly higher performance in the bilateral condition than in the unilateral condition for one out of the five repetitions. It is unclear the extent to which this finding suggests a real underlying effect of bilateral fitting on memory for speech.
Performance in the recall task significantly decreased with increasing age (for the asymptote serial position). A significant benefit of bilateral implantation of 9.7% points was observed, on average, in the listener group up to 65 years old (for the asymptote serial position), while no benefit was obtained in the older listeners. Memory for speech has been previously shown to decrease with age, as reported in several studies (Rabbitt 1991; Anderson et al. 1998; Wingfield et al. 2005; Naveh-Benjamin et al. 2005; Tun et al. 2009). The reduced ability to remember information, as well as other declines in cognitive functions associated with aging, have been suggested to be characterized by a slowing of processing rates, reduced working memory capacity, and generally reduced attentional resources (Naveh-Benjamin et al. 2005; Wingfield et al. 2005; Tun et al. 2009). Recall performance for items in primacy and asymptote serial positions is assumed to reflect retrieval from a long-term storage component, while recall performance for items in recency positions reflects retrieval from a short-term storage component (Ng et al. 2015). The findings of this study showed that aging could explain about 14% of the variance for recall of items in asymptote positions (i.e., when long-term memory was involved). While working memory capacity was not evaluated in this study, differences in storage capacity among the participants might be another factor explaining the large variability in recall performance for primacy and asymptote serial positions, as shown in previous studies (Unsworth 2007; Sarampalis et al. 2009; Ng et al. 2015).
It is interesting that listeners who were implanted simultaneously on the left and right side showed a significantly higher recall performance for the bilateral than for the unilateral condition, with a difference between the two conditions of 5.34% points. However, with only seven listeners implanted simultaneously and an age bias (the simultaneously implanted listeners were younger), it is not possible to disentangle whether this effect can be ascribed to age or simultaneous/sequential implantation (or both). More investigations are needed to rule out (or confirm) an effect of bilateral fitting on memory for speech.
Relationship Between SSQ12 (Spatial Scale) and MAA
Given the accurate performance obtained in the MAA test for listeners fitted with bilateral BAHS, it may seem somewhat surprising that self-reported performance obtained in the SSQ12 questionnaire covered a large range of ratings for the spatial scale (from 0 to 9; see Fig. 6) and was, on average, rather low (3.7; SD = 2.5). However, the three questions of the SSQ12 included in the spatial scale (i.e., locating a dog, judging the distance and direction of a bus) relate to more complex tasks than a simple right versus left discrimination task. As discussed earlier, accurate performance in a right-left discrimination task has been shown to translate to a wide range of sound localization performance in a multi-source task (Grieco-Calub and Litovsky 2010). Hence, the low MAA obtained in this study might not necessarily indicate good localization in everyday life, as reflected by the large range of SSQ12 scores.
The low average score obtained in the spatial scale of the SSQ12 suggests that the bilaterally implanted BAHS listeners included in this study did not feel that they could locate sounds and judge distance and movement very accurately in everyday life. One aspect to consider is that the ratings obtained via the abbreviated version of the SSQ (SSQ12) have been shown to be about 0.6 of a scale point lower than the ratings obtained via the full SSQ questionnaire (SSQ49; Noble et al. 2013). According to the most recent systematic review on bilateral studies with BAHS users (Heath et al. 2022), only one previous study reported SSQ scores for a group of BAHS users (children) that were all bilaterally implanted (Dun et al. 2010). Dun and coauthors reported a score of 5.8 in the spatial scale of the SSQ49 for bilaterally fitted children. Caspers et al. (2022) also collected SSQ scores on bilaterally implanted BAHS users (adults) but the data were not reported. Other studies (Kunst et al. 2008; Bosman et al. 2018; Cuda et al. 2021; Bosman et al. 2021) collected SSQ49 scores using a group of adult users—most of which were unilaterally implanted. The scores obtained by Bosman et al. (2018) for the spatial factor of the SSQ49 (Akeroyd et al. 2014) were very similar to those obtained here (2.6 with an older sound processor and 3.9 with a newer superpower sound processor; 67% of the listeners were unilaterally fitted), as well as the SSQ49 scores obtained in the spatial scale by Cuda et al. (2021) for a group of unilaterally fitted BAHS users (average score of 4.2). Kunst et al. (2008) and Bosman et al. (2021) reported higher mean scores for unilaterally fitted adult BAHS users, 6.8 and 5.1 in the spatial scale, respectively. While there is a high variability in average ratings across studies, there seems to be a general tendency for a rather low self-perceived sound localization performance in everyday life for BAHS users.
Relationship Between SSQ12 and SWIR
A significant correlation was obtained between SSQ scores (total, speech, and qualities) and recall performance in the SWIR test for the bilateral condition. This finding is somewhat in contrast with the outcomes of Ng et al. (2013b), where no correlation was obtained between the two measures. Ng et al. (2013b) found, however, a negative correlation between recall performance in the SWIR test and item 3 of the International Outcome Inventory–Hearing Aids, which reflects remaining hearing difficulties in challenging listening situations. The authors argued that ”the individuals who have better cognitive spare capacity and thus more working memory resources at their disposal… would be better able to engage in explicit kinds of processing in adverse listening situations” and, hence, report more remaining hearing difficulties in challenging situations. At the same time, ”these individuals could achieve better speech understanding and better hearing aid use and thus report greater overall benefit/success.” Hence, it may be not that surprising that a positive correlation was obtained in the present study, where listeners with better recall ability reported to achieve better performance in everyday life. A difference between the present study and the study from Ng et al. (2013b) is the number of sentences in the SWIR test. In Ng et al. (2013b), each list of the SWIR test consisted of eight sentences—which led to a cognitively very demanding test and may have affected the correlation with SSQ (personal communication with Ng). In a recent study using the SWIR test with seven sentences (unpublished data from Ng), a positive correlation was obtained between SSQ scores (full questionnaire with 49 items) and SWIR recall ability, in agreement with the current findings.
The relationship between self-reported performance in everyday life and a test of memory for speech performed in the lab does not seem straightforward, considering that the SSQ does not only focus on hearing difficulties in adverse listening situations, where working memory is used substantially (Ng et al. 2013b). One explanation could be that both the SSQ12 questionnaire and the SWIR test reflect the perceived sound quality of speech. Lunner et al. (2016) used the SWIR test in BAHS users wearing the sound processor mounted on abutment versus softband and concluded that a better sound quality, delivered via the abutment (i.e., a direct-drive solution), was responsible for the increased recall ability. On the other hand, the SSQ12 questionnaire also contains some questions directly targeting sound quality. It may be that those listeners perceiving a good sound quality in real life are also those performing best in the recall task. Possibly supporting this hypothesis, a significant correlation was found only for speech and qualities of hearing scales but not for the spatial scale. Another explanation could be that individual cognitive abilities may affect both outcome measures. In other words, listeners with higher cognitive abilities performed better in the recall task and rated their everyday performance higher. Listeners with good cognitive abilities may be, in fact, more likely to benefit from amplification and advanced signal processing in hearing devices than listeners with poorer cognitive abilities (Akeroyd 2008; Ng et al. 2013a) and, hence, to report better performance.
Finally, age may also have a detrimental effect on both outcome measures. A worsening in recall performance with increasing age was obtained in the present study. Better recall performance in early list positions, which reflects better encoding of words into long-term storage, was shown to be associated with better cognitive capacity (Unsworth 2007). Although SSQ12 scores were not correlated with age in the present study, SSQ scores were shown to be affected by age in a previous study (Banh et al. 2012). Hence, it may be that age (or cognitive factors associated with age) and/or cognitive spare capacity are underlying factor(s) affecting both outcome measures.
Bilateral BAHS Provision
The large improvement in localization acuity obtained in this study indicates that bilaterally fitted BAHS users with symmetrical BC hearing losses can successfully use binaural cues. These findings add to the growing body of evidence reporting the benefits of bilateral BAHS fitting (van der Pouw et al. 1998; Bosman et al. 2001; Dutt et al. 2002; Stenfelt 2005; Ho et al. 2009; Dun et al. 2010; Janssen et al. 2012; Stenfelt and Zeitooni 2013; Dun et al. 2013; Zeitooni et al. 2016; Heath et al. 2022). While providing evidence of bilateral BAHS benefits is needed, other reasons remain for the low penetration of bilateral BAHS implantation, including additional surgical complications and cost-effectiveness of bilateral BAHS provision (Heath et al. 2022). Many centers throughout the world can implant only a certain number of patients per year and, today, several centers choose to implant more patients unilaterally than fewer patients bilaterally. Addressing this knowledge gap together with more evidence on bilateral BAHS benefits might support centers to increase bilateral BAHS provision.
Results from 24 listeners with bilateral conductive or mixed hearing loss showed a large benefit in spatial resolution when being fitted bilaterally (as compared to unilaterally fitted). Performance in the right/left discrimination task was, on average, greater by 20% points at all tested angles when the listeners were fitted with two BAHS than with only one BAHS. In addition, there was no overall benefit of bilateral fitting on memory for speech, despite observing a benefit in one out of five repetitions of the SWIR test. Performance in the SWIR test was also correlated with the users’ self-reported performance in everyday life, such that users with higher recall ability reported to achieve better performance in real life. These findings highlight the advantages of bilateral fitting on spatial resolution, although bilaterally fitted BAHS users continue to experience some difficulties in their daily lives, especially when locating sounds, judging distance and movement. More research is needed to support a higher penetration of bilateral BAHS treatment for bilateral conductive and mixed hearing losses.
This research was funded by Oticon Medical AB, Askim, Sweden. The authors would like to thank Marion Atkin and Roger Esson for their contribution to data collection in this study, and Christian Stender Simonsen from Oticon A/S (Smørum, Denmark) for the technical support.
Abrams H., Kihm J. (2015). An introduction to MarkeTrak IX: A new baseline for the hearing aid market. Hear Rev, 22, 16.
Agterberg M. J., Snik A. F., Van de Goor R. M., Hol M. K., Van Opstal A. J. (2019). Sound-localization performance of patients with single-sided deafness is not improved when listening with a bone-conduction device. Hearing research, 372, 62–68.
Akeroyd M. A. (2008). Are individual differences in speech reception related to individual differences in cognitive ability? A survey of twenty experimental studies with normal and hearing-impaired adults. Int J Audiol, 47(Suppl 2), S53–S71.
Akeroyd M., Guy F., Harrison D., et al. (2014). A factor analysis of the SSQ (speech, spatial, and qualities of hearing scale). Int J Audiol, 53, 101–114.
Anderson N. D., Craik F. I., Naveh-Benjamin M. (1998). Attentional demands of encoding and retrieval in younger and older adults: I. Evidence from divided attention costs. Psychol Aging, 13, 405–423.
Ashmead D. H., Clifton R. K., Perris E. E. (1987). Precision of auditory localization in human infants. Dev Psychol, 23, 641–647.
Baguley D. M., Bird J., Humphriss R. L., Prevost A. T. (2006). The evidence base for the application of contralateral bone anchored hearing aids in acquired unilateral sensorineural hearing loss in adults. Clin Otolaryngol, 31, 6–14.
Banh J., Singh G., Pichora-Fuller M. K. (2012). Age affects responses on the Speech, Spatial, and Qualities of Hearing Scale (SSQ) by adults with minimal audiometric loss. J Am Acad Audiol, 23, 81–91; quiz 139.
Bartzokis G. (2004). Age-related myelin breakdown: A developmental model of cognitive decline and Alzheimer’s disease. Neurobiol Aging, 25, 5–18; author reply 49.
Bench J., Kowal A., Bamford J. (1979). The BKB (Bamford-Kowal-Bench) sentence lists for partially-hearing children. Br J Audiol, 13, 108–112.
Bernstein L. R. (2004). Sensitivity to interaural intensitive disparities: Listeners’ use of potential cues. J Acoust Soc Am, 115, 3156–3160.
Blauert J (1997) Spatial Hearing: The psychophysics of human sound localization.: MIT Press.
Bosman A. J., Snik A. F. M., Van der Pouw C. T. M., Mylanus E. A., Cremers C. W. (2001). Audiolometric evaluation of bilaterally fitted bone-anchored hearing aids. Audiology, 40, 158–167.
Bosman A. J., Kruyt I. J., Mylanus E. A. M., Hol M. K. S., Snik A. F. M. (2018). Evaluation of an abutment-level superpower sound processor for bone-anchored hearing. Clinical Otolaryngology, 43, 1019–1024.
Bosman A. J., Christensen J. H., Rosenbom T., et al. (2021). Investigating real-world benefits of high-frequency gain in bone-anchored users with ecological momentary assessment and real-time data logging. J Clin Med, 10, 3923.
Caspers C. J., Janssen A. M., Agterberg M. J. H., Cremers C. W. R. J., Hol M. K. S., Bosman A. J. (2022). Sound localization with bilateral bone conduction devices. European Archives of Oto-Rhino-Laryngology, 279, 1751–1764.
Cochlear bilateral market survey (2018) Bilateral benefits of the Baha System: A summary of clinical evidence. Report Number BUN713 ISS1 DEC18. December 2018. www.Cochlear.com/US
Cuda D., Murri A., Mochi P., Mainardi A. (2021). Postoperative benefit of bone anchored hearing systems: Behavioral performance and self-reported outcomes. Int Arch Otorhinolaryngol, 26, e314–e320.
Dun C. A., de Wolf M. J., Mylanus E. A., Snik A. F., Hol M. K. S., Cremers C. W. R. J. (2010). Bilateral bone-anchored hearing aid application in children: The Nijmegen experience from 1996 to 2008. Otol Neurotol, 31, 615–623.
Dun C. A., Agterberg M. J., Cremers C. W., Hol M. K. S., Snik A. F. M. (2013). Bilateral bone conduction devices: Improved hearing ability in children with bilateral conductive hearing loss. Ear Hear, 34, 806–808.
Dutt S. N., McDermott A. L., Burrell S. P., Cooper H. R., Reid A. P., Proops D. W. (2002). Patient satisfaction with bilateral bone-anchored hearing aids: the Birmingham experience. J Laryngol Otol Suppl, 116, 37–46.
Grieco-Calub T. M., Litovsky R. Y. (2010). Sound localization skills in children who use bilateral cochlear implants and in children with normal acoustic hearing. Ear Hear, 31, 645–656.
Heath E., Dawoud M. M., Stavrakas M., Ray J. (2022). The outcomes of bilateral bone conduction hearing devices (BCHD) implantation in the treatment of hearing loss: A systematic review. Cochlear Implants International, 23, 95–108.
Ho E. C., Monksfield P., Egan E., Reid A., Proops D. (2009). Bilateral bone-anchored hearing aid: Impact on quality of life measured with the Glasgow benefit inventory. Otol Neurotol, 30, 891–896.
Hofman P. M., Van Opstal A. J. (1998). Spectro-temporal factors in two-dimensional human sound localization. J Acoust Soc Am, 103, 2634–2648.
Janssen R. M., Hong P., Chadha N. K. (2012). Bilateral bone-anchored hearing aids for bilateral permanent conductive hearing loss: A systematic review. Otolaryngol Head Neck Surg, 147, 412–422.
Kochkin S. (2009). MarkeTrak VIII: 25-year trends in the hearing health market. Hear Rev, 16, 12–31.
Kunst S. J., Hol M. K., Mylanus E. A., Leijendeckers J. M., Snik A. F. M., Cremers C. W. R. J. (2008). Subjective benefit after BAHA system application in patients with congenital unilateral conductive hearing impairment. Otol Neurotol, 29, 353–358.
Kuznetsova A., Brockhoff P. B., Christensen R. H. (2017). lmerTest package: Tests in linear mixed effects models. J Stat Softw, 82, 1–26.
Lagerkvist H., Carvalho K., Holmberg M., Petersson U., Cremers C., Hultcrantz M. (2020). Ten years of experience with the Ponto bone-anchored hearing system—A systematic literature review. Clin Otolaryngol, 45, 667–680.
Levitt H. (1971). Transformed up-down methods in psychoacoustics. J Acoust Soc Am, 49, 467–477.
Litovsky R. Y. (1997). Developmental changes in the precedence effect: Estimates of minimum audible angle
. J Acoust Soc Am, 102, 1739–1745.
Litovsky R. Y., Johnstone P. M., Godar S., Agrawal S., Parkinson A., Peters R., Lake J. (2006). Bilateral cochlear implants in children: Localization acuity
measured with minimum audible angle
. Ear Hear, 27, 43–59.
Lunner T. (2003). Cognitive function in relation to hearing aid use. Int J Audiol, 42, S49–S58.
Lunner T., Rudner M., Rosenbom T., Ågren J., Ng E. H. N. (2016). Using speech recall in hearing aid fitting and outcome evaluation under ecological test conditions. Ear Hear, 37, 145S–154S.
Macleod A., Summerfield Q. (1990). A procedure for measuring auditory and audiovisual speech-reception thresholds for sentences in noise: Rationale, evaluation, and recommendations for use. Br J Audiol, 24, 29–43.
Merat N., Groeger J. A. (2003). Working-memory and auditory localization: Demand for central resources impairs performance. Q J Exp Psychol, 56, 531–549.
Mills A. W. (1958). On the minimum audible angle
. J Acoust Soc Am, 30, 237–246.
Naveh-Benjamin M., Craik F., Guez J., et al. (2005). Divided attention in younger and older adults: Effects of strategy and relatedness on memory performance and secondary task costs. Journal of Experimental Psychology: Learning.
Mem Cognit, 31, 520–537.
Neher T., Laugesen S., Søgaard Jensen N., et al. (2011). Can basic auditory and cognitive measures predict hearing-impaired listeners’ localization and spatial speech recognition abilities? J Acoust Soc Am, 130, 1542–1558.
Neher T., Lunner T., Hopkins K., Moore B. C. J. (2012). Binaural temporal fine structure sensitivity, cognitive function, and spatial speech recognition of hearing-impaired listeners (L). J Acoust Soc Am, 131, 2561–2564.
Ng E. H., Rudner M., Lunner T., Pedersen M. S., Rönnberg J. (2013a). a) Effects of noise and working memory capacity on memory processing of speech for hearing-aid users. Int J Audiol, 52, 433–441.
Ng E. H., Rudner M., Lunner T., Rönnberg J. (2013b). b) Relationships between self-report and cognitive measures of hearing aid outcome. Speech Lang Hear, 16, 197–207.
Ng E., Rudner M., Lunner T., et al. (2015). Noise reduction improves memory for target language speech in competing native but not foreign language speech. Ear Hear, 36, 82–91.
Noble W., Jensen N. S., Naylor G., Bhullar N., Akeroyd M. A. (2013). A short form of the speech, spatial and qualities of hearing scale suitable for clinical use: The SSQ12. Int J Audiol, 52, 409–412.
Pichora-Fuller M. K., Schneider B. A., Daneman M. (1995). How young and old adults listen to and remember speech in noise. J Acoust Soc Am, 97, 593–608.
van der Pouw K. T., Snik A. F., Cremers C. W. (1998). Audiometric results of bilateral bone-anchored hearing aid application in patients with bilateral congenital aural atresia. Laryngoscope, 108(4 Pt 1), 548–553.
Priwin C., Stenfelt S., Granstrom G., Tjellström A., Håkansson B. (2004). Bilateral bone-anchored hearing aids (BAHAs): An audiometric evaluation. Laryngoscope, 114, 77–84.
Rabbitt P. (1991). Mild hearing loss can cause apparent memory failures which increase with age and reduce with IQ. Acta Otolaryngol, 11(Supp 476), 167–176.
Ries D. T., Hamilton T. R., Grossmann A. J. (2010). The effects of intervening interference on working memory for sound location as a function of inter-comparison interval. Hear Res, 268, 227–233.
Rönnberg J., Rudner M., Foo C., et al. (2008). Cognition counts: A working memory system for ease of language understanding (ELU). Int J Audiol, 47, S99–S105.
Sarampalis A., Kalluri S., Edwards B., Hafter E. (2009). Objective measures of listening effort: Effects of background noise and noise reduction. J Speech Lang Hear Res, 52, 1230–1240.
Smeds K., Wolters F., Rung M. (2015). Estimation of signal-to-noise ratios in realistic sound scenarios. J Am Acad Audiol, 26, 183–196.
Snik A. F., Mylanus E. A., Cremers C. W. (1995). Bone-anchored hearing aids in patients with sensorineural hearing loss and persistent otitis externa. Clin Otolaryngol Allied Sci, 20, 31–35.
Snik A. F., Bosman A. J., Mylanus E. A., Cremers C. W. R. J. (2004). Candidacy for the bone-anchored hearing aid. Audiol Neurootol, 9, 190–196.
Snik A., Maier H., Hodgetts B., Kompis M., Mertens G., van de Heyning P., Lenarz T., Bosman A. (2019). Efficacy of auditory implants for patients with conductive and mixed hearing loss depends on implant center. Otol Neurotol, 40, 430–435.
Stenfelt S. (2005). Bilateral fitting
of BAHAs and BAHA fitted in unilateral deaf persons: Acoustical aspects. Int J Audiol, 44, 178–189.
Stenfelt S. (2012). Transcranial attenuation of bone-conducted sound when stimulation is at the mastoid and at the bone conduction hearing aid position. Otol Neurotol, 33, 105–114.
Stenfelt S., Zeitooni M. (2013). Binaural hearing ability with mastoid applied bilateral bone conduction stimulation in normal hearing subjects. J Acoust Soc Am, 134, 481–493.
Tjellstrom A., Hakansson B. (1995). The bone-anchored hearing aid. Design principles, indications, and long-term clinical results. Otolaryngol Clin North Am, 28, 53–72.
Tun P. A., McCoy S., Wingfield A. (2009). Aging, hearing acuity, and the attentional costs of effortful listening. Psychol Aging, 24, 761–766.
Unsworth N. (2007). Individual differences in working memory capacity and episodic retrieval: Examining the dynamics of delayed and continuous distractor free recall. J Exp Psychol Learn Mem Cogn, 33, 1020–1034.
Van Wanrooij M. M., Van Opstal A. J. (2007). Sound localization under perturbed binaural hearing. J Neurophysiol, 97, 715–726.
Westover L., Ostevik A., Aalto D., Cummine J., Hodgetts W. E. (2020). Evaluation of word recognition and word recall with bone conduction devices: Do directional microphones free up cognitive resources? Int J Audiol, 59, 367–373.
Wingfield A., Tun P. A., McCoy S. L. (2005). Hearing loss in older adulthood: What it is and how it interacts with cognitive performance. Curr Dir Psycholog Sci, 14, 144–148.
Zeitooni M., Maki-Torkko E., Stenfelt S. (2016). Binaural hearing ability with bilateral bone conduction stimulation in subjects with normal hearing: Implications for bone conduction hearing aids. Ear Hear, 37, 690–702.